[erlang-questions] light-weight concurrency in C

[Ludovic Coquelle]

I haven't done any serious testing or benchmarking, but I'm guessing that
each servers has its own advantages depending of use cases.
With a slow network, Yaws can be a good solution as it should be able to
handle a lot of concurrent connections even if its efficiency for each
request is a bit less. Am I wrong?

On 6/21/07, Tom Samplonius <tom@REDACTED> wrote:
>
>
> ----- "Bob Ippolito" <bob@REDACTED> wrote:
> > On 6/18/07, Serge Aleynikov <saleyn@REDACTED> wrote:
> > > This looks like the same approach to concurrency as what's been used
> > in
> > > Erlang: http://legacy.imatix.com/html/smt.  They use it in their
> > open
> > > source high-performance messaging server: http://www.openamq.org.
> > >
> > > While there's been a comparison between Yaws and Apache, it would
> > be
> > > interesting to see one against http://www.xitami.com as it would
> > seem
> > > like it could actually outperform Yaws.
> >
> > I also expect nginx and lighttpd to be faster than yaws. In my
> > informal benchmarks nginx is about twice as fast as an erlang-based
> > httpd that I wrote... and I'm pretty sure my http server is faster
> > than yaws because it does less work.
> >
> > -bob
>
>
>   I don't think is surprising that nginx and lighttpd are faster than
> Yaws.  erlang is not really a language to just push bits from disk to
> network.  It doesn't seem optimized for it either, either the runtime or the
> language.  erlang doesn't have a lot of low-level IO primitives to really
> push bits fast.  And it probably won't, since erlang is more portable than
> nginx and lighttpd.
>
>
> Tom
> _______________________________________________
> erlang-questions mailing list
> erlang-questions@REDACTED
> http://www.erlang.org/mailman/listinfo/erlang-questions
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20070621/bb00abf4/attachment.htm>